I respect your interpretation, what I respectfully have to be critical of is your rationale (aside from your telling me what I should or shouldn’t do).

First, to the “five factors” you mention. Your statement, “if you look at how they decided on the five factors they ended up with…”

Those five factors are the well-known lifestyle habits that have been shown again and again to be the most prominent causes of lifestyle-dependent metabolic disease: smoking, physical inactivity, poor diet, heavy alcohol intake, and overweight. The authors looked at how different combinations of those factors affect life and health expectancy. There is nothing wrong with the choice of those 5, and certainly nothing manipulative. In fact, I have addressed the issue of non-predicatibility in an earlier post: https://medium.com/right-to-rejuvenation/cvd-risk-factors-may-actually-not-kill-you-or-make-you-sick-the-stunning-statistics-e1c835d30f6a

Second, to the “model” you mention: “…they just kept adding factors to their model until they reached what they considered a significant increase in life expectancy.” That is a misinterpretation of what the authors say and of the statistical method of covariate adjustment. The five factors were the independent predictor variables. The authors adjusted them for covariates to improve the signal-to-noise ratio in the data. Adjustment for covariates is not something you do to “massage” data to deliver the desired results. It is something you have to do to uncover true relationships. In this case the authors adjusted for seven and nine covariates for men and women respectively. Among those covariates were family history of diabetes, cancer and cardiovascular disease, as well as menopausal status and hormone therapy. These variables are well-known correlates of life and health expectancy, and they will pollute any data analysis if not adjusted for.

If I may make a concluding remark and suggestion to this discussion: Feel free to write an article about your views and interpretation.